Re/AC catalysts for selective hydrogenation of dimethyl 1, 4-cyclohexanedicarboxylate to 1, 4-cyclohexanedimethanol: Essential roles of metal dispersion and chemical environment

Article abstract of DOI:10.1016/j.apcata.2020.117669

Rhenium, although viewed as one of the noble metals with lower-price, has been commonly used as doping element in the bimetallic catalysts due to its middlebrow to activate hydrogen. Its major role as catalyst is less frequently mentioned. In this work, rhenium has been decorated on the surface of activated carbon and used for the selective hydrogenation of dimethyl 1, 4-cyclohexanedicarboxylate (DMCD) to 1, 4-cyclohexanedimethanol (CHDM). Characterizations suggested that ReO_x particles were anchored occupying the surface oxygenated groups on pre-functionalized carbon. Rhenium decoration modified both the textural and chemical properties of the samples. Electrons were easily transferred from Re to the neighboring C atoms as a result of the formation of fine ReO_x particles. Medium strong acid sites were generated and rhenium species in the reduced states could be still maintained under appropriate rhenium dispersion. The moderate hydrogenation ability of rhenium catalyst partially restrained the excessive hydrogenation of CHDM to other by-products. Rational decoration of 5 wt% Re performed the better catalytic performance with complete conversion of diester and 66 % yield of diol. The specific rate reached 9.5×10² mmol_DMCD?g_Re^-1?h^-1 at 220 °C under 10 MPa H₂.

Full text of DOI:10.1016/j.apcata.2020.117669

Journal Pre-proof
Re/AC catalysts for selective hydrogenation of dimethyl 1,
4
-cyclohexanedicarboxylate to 1, 4-cyclohexanedimethanol: Essential
roles of metal dispersion and chemical environment
Jingjie Luo (Conceptualization) (Investigation) (Writing - original
draft), Enhui Qu (Data curation) (Investigation), Yixue Zhou
(
(
(
Validation), Yanan Dong (Formal analysis), Changhai Liang
Writing - review and editing) (Supervision) (Project administration)
Funding acquisition)
PII:
S0926-860X(20)30262-3
DOI:
https://doi.org/10.1016/j.apcata.2020.117669
APCATA 117669
Reference:
To appear in:
Applied Catalysis A, General
Received Date:
Revised Date:
Accepted Date:
2 March 2020
15 April 2020
27 May 2020
Please cite this article as: { doi: https://doi.org/
This is a PDF file of an article that has undergone enhancements after acceptance, such as
the addition of a cover page and metadata, and formatting for readability, but it is not yet the
definitive version of record. This version will undergo additional copyediting, typesetting and
review before it is published in its final form, but we are providing this version to give early
visibility of the article. Please note that, during the production process, errors may be
discovered which could affect the content, and all legal disclaimers that apply to the journal
pertain.
©
2020 Published by Elsevier.

Re/AC catalysts for selective hydrogenation of dimethyl 1, 4-
cyclohexanedicarboxylate to 1, 4-cyclohexanedimethanol:
Essential roles of metal dispersion and chemical environment
Jingjie Luo, Enhui Qu, Yixue Zhou, Yanan Dong, Changhai Liang*
State Key Laboratory of Fine Chemicals and Laboratory of Advanced Materials & Catalytic
Engineering (AMCE), School of Chemical Engineering, Dalian University of Technology,
Panjin Campus, Panjin 124221, China.
*
Corresponding author, E-mail: changhai@dlut.edu.cn (CH Liang)
Graphical Abstract
1

Highlights





Pre-functionalized Re /AC are used for selective hydrogenation of diester to diol.
x
Higher dispersion of exposed Re sites changes the textural and redox property.
Electron is delivered from Re to C atoms leading to electron-deficient Re sites.
New medium strong acid sites are generated due to fine ReO
x
particles.
2
o
Re
5
/AC displays specific rate of 9.5 × 10 mmol ∙ ꢀ 푅− 푒ꢁ ∙ ℎ−ꢁ at 220 C in 10 MPa H
2
.
Abstract
Rhenium, although viewed as one of the noble metals with lower-price, has been
commonly used as doping element in the bimetallic catalysts due to its middlebrow to activate
hydrogen. Its major role as catalyst is less frequently mentioned. In this work, rhenium has
been decorated on the surface of activated carbon and used for the selective hydrogenation of
dimethyl 1, 4-cyclohexanedicarboxylate (DMCD) to 1, 4-cyclohexanedimethanol (CHDM).
Characterizations suggested that ReO particles were anchored occupying the surface
x
oxygenated groups on pre-functionalized carbon. Rhenium decoration modified both the
textural and chemical properties of the samples. Electrons were easily transferred from Re to
the neighboring C atoms as a result of the formation of fine ReO particles. Medium strong
x
acid sites were generated and rhenium species in the reduced states could be still maintained
under appropriate rhenium dispersion. The moderate hydrogenation ability of rhenium catalyst
partially restrained the excessive hydrogenation of CHDM to other by-products. Rational
decoration of 5 wt% Re performed the better catalytic performance with complete conversion
2

of diester and 66% yield of diol. The specific rate reached 9.5 × 10 푚푚표푙퐷푀퐶퐷 ∙ ꢀ ∙ ℎ−ꢁ
2
−ꢁ
푅푒
o
at 220 C under 10 MPa H
2
.
Keywords: Rhenium catalyst, Selective hydrogenation, DMCD, CHDM, Activated carbon,
Surface functionalization
1
. Introduction
1
, 4-cyclohexanedimethanol (CHDM) is a highly value-added organic compound
extensively used in the polymer industry [1, 2]. It is noteworthy that CHDM is now preferred
over ethylene glycol as a stepping-stone to produce polyester fibers [3, 4]. The most feasible
catalytic route to date for producing CHDM is still a two-step process, via firstly the
hydrogenation of dimethyl terephthalate (DMT) to dimethyl 1, 4-cyclohexanedicarboxylate
o
(
DMCD), and subsequently hydrogenated to CHDM in the temperature range of 200-300 C
under 3-15 MPa H [5-7]. There are also successful one-pot synthesis of CHDM directly from
2
DMT, unfortunately in the cost of lower conversion and turnover frequency, or the excessive
loading of precious metals [7, 8]. No matter how on an industrial scale, the most feasible route
to CHDM is the catalytic hydrogenation of DMCD by using Cu-Cr based catalysts. The
toxicity of such additional Cr species leads to severe environmental problems, Cr-free catalysts
with high efficiency are urgently to be designed.
In recent study, Hu et al. [9] have synthesized a series of Cu-Zn-Al catalysts used for the
fixed-bed hydrogenation of DMCD to CHDM. The reduction process and the stably dispersed
+
metallic Cu and Cu species were confirmed to be a key point to activate the ester groups of
3

reactant. However, the reaction conditions were relative harsh under high pressure with huge
/DMCD ratio (406 mol/mol) and large dosage of catalyst (2.0-3.0 g). Other Ru or Pt based
H
2
catalysts are also used as main catalysts for the production of CHDM, and bimetallic or even
trimetallic catalysts have been designed for the reaction. The catalytic performance by using
such multicomponent catalyst seems even mysterious to be understood, not to mention their
unsatisfied activity. Besides, there is in fact not much work reported concerning the
hydrogenation of ester groups of DMCD to CHDM. It is still of great importance to find new
catalysts that are available to catalyze the hydrogenation of esters with well-established
structure-performance relationship of catalysts.
Till now, precious metal based catalysts have displayed satisfied hydrogenation
performance for many heterogeneous reaction [10, 11]. Different from other metals, Re
precursor has the much lower price but inferior hydrogenolysis ability compared to other
precious metals. It was reported that the additional Re species in precious metal based catalyst
effectively adjust the dehydrogenation and C-O dissociation ability of active metal sites,
remitting the excessive production of by-products [12-14]. Our previous work also suggested
that using Re as dopant for the precious metal catalyst, fine distribution and smaller average
sizes could be achieved [15, 16]. However until now, there is rarely report concerning the
hydrogenation ability and the surface chemical environment of monometallic Re catalyst to
understand the selective hydrogenation process of DMCD.
Under such circumstance, a series of ReO nanoparticles have been supported on activated
x
carbon with different pre-treatments. The influence of Re decoration and pre-treatment on the
4

surface chemical composition, metal dispersion, and their rational performances on the
selective hydrogenation ability have been studied in detail.
2
. Experimental part
2
.1 Materials
All chemicals and reagents were of analytical grade. The NH
4
ReO and activated carbon
4
were purchased from Aladdin Chemical Co., Ltd. All the other chemical reagents such as the
7% H and concentrated HNO employed in this study were obtained from Sinopharm
chemical reagent Co. Ltd. Deionized (DI) water was used in the experiments.
3
2
O
2
3
2
.2 Catalyst preparation
Activated carbon was preliminarily washed with boiled water to remove the residual ash
content. It was previously reported that porous structure and the surface area, as well as the
surface chemical environment might be modified in case of pre-treatments [17]. Herein,
different pretreatments were used to modify the activated carbon: i) activated carbon was
o
thermally treated under pure H
was treated by ii) 37% H
h, denoted as AC-HO and AC-N, respectively; vi) thermally treated by H
by concentrated HNO solution for 5 h, denoted simply as AC. Wash and filter of samples
2
at 800 C for 10 h, and denoted as AC-H; activated carbon
2
O
2
aqueous solution for 5 h or iii) concentrated HNO
3
solution for 5
2
and subsequently
3
were necessary until the pH value of solution reaches 7.0.
Re based catalysts have been prepared by incipient wetness impregnation. Certain amount
of NH
4
ReO
4
as precursor was dissolved in distilled water. 1g activated carbon was added into
5

the aqueous solution of Re precursor with stirring and sonicated for 30 min. The mixture was
o
aged for 16 h and dried directly overnight in an oven under 60 C. The as-obtained dry powder
is label as Re /AC-y (x=1, 3, 5, 7, 9; y suggests the type of pretreatment).
x
2
.3 Characterization
Transmission electron microscopy (TEM) was performed on a FEI Tecnai F30 electron
microscope at an acceleration voltage of 120 kV. Before each observation, the catalyst was
dissolved in solvent and treated in sonication for 10 min. Several drops of the supernatant
liquid were added and evaporated on the copper grid. BET surface area and the pore structure
of the catalysts were investigated using N physical adsorption/desorption at -196 °C on an
2
automatic volumetric sorption analyzer (Quantachrome, Autosorb-iQ-C). The specific surface
area was calculated by the BET equation and the pore size was derived from the desorption
branch of the isotherms by BJH method. Temperature-programmed reactions were operated
on a Quantachrome ChemStar3000 equipment. H
2
-temperature-programmed reduction (H -
2
TPR) was tested using 100 mg sample. The sample was purged in Argon for 2 h at room
temperature. After the gas atmosphere was changed to the mixture of 10% hydrogen in argon,
o
the reduction data was recorded in the range of 50-800 C. During the ammonia temperature-
programmed desorption (NH
3
-TPD), reduced catalyst (100 mg) was pre-treated by NH /He
3
o
gas mixture at 50 C for 1 h. After purged with He, the desorption was carried out and recorded
o
o
o
from 50 to 700 C. X-ray diffraction (XRD) measurements were carried out within 10 -80
on the Shimadzu XRD-7000S diffractometer with a Cu Kαradiation at 40 kV and 30 mA. X-
ray photoelectron spectroscopy (XPS) measurements were performed on an ESCALAB™
6

2
50Xi (Thermo Fisher, USA) with the monochromatic Al Kα (1486.6 eV) X-ray source. CO
chemisorption was carried out using a Micrometer Autochem 2910 instrument. The Raman
spectra was recorded using a 532 nm laser source by Renishaw’s in Via Confocal Raman
Microscope. The surface defect density was evaluated according to the equation proposed by
Cançado et al. [18].
푰
푫
−
ퟐ
ퟗ
ퟒ
풏 (풄풎 ) = (ퟕ. ퟑ ± ퟐ. ퟐ) × ퟏퟎ 푬 ꢂ ꢃ
푫
푳
푰
푮
where I /I
D G
is the ratio of peak intensity by height in Raman spectra locating around 1340 cm-
1
-1
and 1570 cm . E
L
is the excitation energy around 2.41 eV for single layer and few layer
graphene [19].
o
Before each characterization, the catalyst was reduced under H
2
/Ar atmosphere at 460 C
for 3 h, to ensure their identical condition as those before the reaction test, except the H
2
-TPR.
2
.4 Catalyst Evaluation
Liquid-phase hydrogenation was performed in 50 mL autoclave with electric temperature
controller. Certain amount of reactant dimethyl 1, 4-cyclohexanedicarboxylate (DMCD) was
o
dissolved in isopropanol with 125 mg catalyst pre-reduced by H
2
/Ar under 460 C for 3 h. The
reactor was purged with hydrogen for three times to remove residual air and the aimed reaction
pressure was set. The reactor was heated subsequently to the desired temperature under
constant stirring. After reaction, the reactor was cooled down to room temperature. The liquid
samples were analyzed by gas chromatograph (GC-7890F, FID, FFAP column 30 m×0.32 mm
×0.5 mm) using the P-xylene as internal standard.
7

The conversion of DMCD (푥퐷푀퐶퐷) , selectivity (푆 ) and the relative yield (푌ꢄ) of
푖
different products were calculated by the following equations [20]:
푛ꢅ ꢆ 푛푡
Conversion(푥퐷푀퐶퐷) =
× 100%
푛ꢅ
푛푖
Selectivity(푆 ) =
× 100%
푖
∑
푛푖
Yield(푌ꢄ) = 푥퐷푀퐶퐷 × 푆 × 100%
푖
Where n
0
represents the initial number of moles of DMCD and n represents the number of
t
moles of DMCD at time t. n represents the molar quantity of product i.
i
The specific reaction rate to convert one mmol of DMCD, based on per gram of surface
exposed active metal per hour, was evaluated by the following equation:
푛푐ꢇꢈ푣. ∙ 10−3
R =
푚푚표푙 ∙ ꢀ−ꢁ ∙ ℎ−ꢁ
푅푒
푚푅푒 ∙ ꢉ푑푖푠푝. ∙ ꢊ
Where 푛푐ꢇꢈ푣. is the number of moles of converted DMCD, in mmol; 푚푅푒 is the mass of Re
species in use, in gram; ꢉ푑푖푠푝. is the surface dispersion of active metal sites tested by CO
chemisorption; and t is the reaction time, in hour.
3
. Results and discussion
3
.1 Chemical composition and texture properties of Re /AC
x
The morphology of the reduced sample is revealed by TEM images as shown in Figure
A. The black spots in each sample should be Re-containing particles due to the big atomic
1
number of Re compared to C. Only very few amount of particles can be observed in the
Re /AC sample, due to the lower Re content. The average particle sizes of Re /AC, Re /AC,
and Re /AC samples are calculated to be 2.1, 3.7, and 5.3 nm, respectively, based on the
1
3
5
9
8

statistical results of more than 200 particles. Re species are easily dispersed on the surface of
activated carbon without obvious sintering of huge particles. As the amount of rhenium
increases, both small and bigger particles appear mainly caused by the slight accumulation of
Re element, especially in the Re /AC sample.
9
Figure 1. TEM images (A), N
2
adsorption/desorption isotherms (B) and H -TPR spectra (C)
2
of the reduced Re /AC samples.
x
The influence of Re species on the structure of activated carbon has been detected by the
adsorption/desorption isotherms (Figure 1B). The BET surface areas and the average pore
N
2
sizes of different samples are listed in Table 1. The tested samples display the typical isotherms
of type I (<0.02) with microspore filling and type IV by capillary condensation at high pressure,
suggesting the co-presence of micro- and meso-pores [21, 22]. The H4 hysteresis loop exists
at relative pressure (P/P
0
) of 0.5-1.0 [23], mainly due to the inter-particle textural pores in the
2
carbon materials. The surface areas of the Re
x
/AC samples change from 541 to 370 m /g as
the Re content increases to 7-9 wt%. The average pore sizes also slightly decrease from 4.6
2
nm to 4.2 nm. In fact, the initial surface area of activated carbon is 486 m /g and the average
pore size is 4.6 nm. The decoration of Re causes the incrementing of the surface area without
9

great change on the pore size, especially if the Re content was lower (e.g. Re
It is inferred that the additional doping and calcination processes in the Re
1
/AC and Re
3
/AC).
x
/AC samples may
result in the rough surface of carbon, which contributes to the larger surface area in turn.
However, as the Re content raises, the metal particles are accumulated on the surface of carbon
and may partially occupy the pores accounting for the loss of both the surface area and pore
sizes.
Table 1 Textural property and surface information of different Re based catalysts.
BET
Metallic
uptakea surface
Specific rate
Surface
Pore
size
(nm)
4.6
Defect density
H
2
Metal
disper.b
2
Sample
(× 10
ꢁ
ꢁ
-2
area
(× 10 cm )
(µmol/kg)
area
mmolDMCD/gRe/s)c
2
2
b
(
m /g)
(m /g)
Re
Re
Re
Re
Re
Re
Re
Re
1
3
5
7
9
5
5
5
/AC
/AC
/AC
/AC
/AC
541
531
554
450
370
5.15
4.88
4.90
4.73
4.65
4.85
4.90
4.83
17.3
25.2
31.4
34.2
40.2
-
21.1
28.3
21.8
30.5
40.1
11.4
22.0
5.2
10.0%
13.5%
6.2%
11.9
7.0
7.7
2.3
1.3
5.5
4.0
2.5
4.6
4.4
4.3
14.5%
19.1%
6.4%
4.2
/AC-HO 642
3.3
/AC-N
588
612
3.8
-
10.5%
2.4%
/AC-H
4.7
-
a
H
2
consumption of α peak in H
2
-TPR evaluated after peak deconvolution.
b
c
Metallic surface area and metal dispersion evaluated by CO chemisorption.
o
Specific rates of Re
x
/AC samples are calculated under reaction condition at 220 C in 8 MPa H , and
2
o
specific rates of Re
5
/AC-y samples are at 220 C in 10 MPa H
2
, weight ratio of DMCD/catalyst=16:1.
The reducibility and redox behavior of different samples are studied by H
2
-TPR (Figure
o
o
1
C). Two reduction peaks locating at 300-400 C and 600 C can be observed in the samples.
The reduction peak at lower temperature (denoted as α peak) should be ascribed to the surface
or subsurface metal species in oxidization states [24-26], and the peak at higher temperature
is assigned to the methanation of carbon with H
2
to form CH
4
[27]. The desorption temperature
/AC is maximized
of α peak varies as a function of Re contents. The α reduction peak in Re
1
1
0

o
at 390 C. Moderate raising of the Re contents (Re
5
/AC and Re /AC samples) results in the
7
o
reduction peak towards lower temperature (310 C) with much intensified peak heights. The
o
further increase of Re content to 9 wt% leads to slight delay of the α peak to 340 C. For better
comparison, the reduction peaks have been deconvoluted and the H
2
consumption of the α
consumption of the α
peak is evaluated as listed in Table 1. Both the intensity and related H
2
peak clearly raises with the metal loading. However, it is worth noting that although the Re
loading increments in different Re /AC samples, the area of the α peak does not change
proportionally. The H consumption by Re /AC is only about 2.3-fold higher than the Re /AC
x
2
9
1
sample. It is inferred that the metal surface dispersion or even the Re-C interaction might be
distinctive by Re decoration in different samples. No matter how, appropriate amount of Re
on activated carbon enhances the reducibility of metal precursor, and the rhenium species
could be further modulated after the H thermal treatment.
2
1
1

Figure 2. XPS spectra in the Re4f region of different reduced samples: Re
C), and Re /AC samples (D).
3
/AC (A), Re
5
/AC (B), Re /AC
7
(
9
The detailed information of surface metal composition in the reduced catalyst is further
revealed by XPS spectra in the Re4f region. The overlapped peaks of rhenium species can be
deconvoluted into two sets of peaks due to a doublet Pd 4f7/2 and 4f5/2 signals. The binding
energy in the Re 4f7/2 region at about 40.8 eV, 42.4 eV, 45.4 eV should be related to the Re0,
4
+
6+
Re , and Re species [28, 29], respectively. Multiple chemical valences are revealed in
6
+
o
different samples with Re species as the main composition even after H
2
reduction at 460 C.
6
+
4+
In the Re /AC sample, 69% Re species exists together with small amount of Re (16%) and
3
0
4+
0
Re (15%), as listed in Table 2. Even lower amounts of Re (12%) and Re (12%) species are
4
+
0
detected in Re
5
/AC. While the Re and Re species are barely observed in Re
9
/AC. It is
suggested by the XPS data that small amount of Re decoration (≤ 5 wt%) facilitates the
1
2

4
+
0
6+
generation of Re and Re in the reduced states. The binding energy of Re species shifts
simultaneously from 45.2 eV of Re /AC to 45.4 eV of Re /AC. Electron transfer is likely to
3
9
be happened from Re to C, which should be mainly resulted from the enhanced Re-C
interaction with rhenium decoration. The surface molar ratio between Re and C is also given
in Table 2. The slightly higher ratio of Re/C, about 1.3-2.0 times compared to the theoretical
value is consistent with the well dispersion of metal species on the surfaces of carbon.
Table 2 Surface composition of different Re catalysts based on XPS spectra.
Sample
Surface Re species
Surface Re Surface oxygenated species
Re6+
Re4+
Re0
O-C=O
C=O
O/C%
dispersion C-O(H)
Re
Re
Re
Re
Re
Re
Re
3
/AC
/AC
/AC
/AC
69%
76%
97%
100%
16%
12%
1%
-
15%
12%
2%
-
0.042
0.063
0.114
0.169
0.053
0.090
0.039
42%
40%
27%
20%
31%
25%
22%
36%
28%
26%
24%
31%
33%
51%
22%
32%
47%
56%
35%
42%
27%
7.2
8.9
7.8
9.3
7.2
7.3
8.1
5
7
9
5
5
5
/AC-HO 93%
3%
-
4%
-
/AC-N
/AC-H
100%
65%
16%
19%
3
.2 Selective hydrogenation from diester to diol
The as-synthesized Re
dimethyl 1,4-cyclohexanedicarboxylate (DMCD, i.e. diester) to 1, 4-cyclohexanedimethanol
CHDM, i.e. diol) (Table 3). Pure activated carbon did not perform any activity for the
x
/AC catalysts have been used for the selective hydrogenation of
(
hydrogenation. Diverse by-products have been detected during the reaction by different
catalysts as listed in Scheme S1 (Supporting information). The target product-CHDM is
transformed preferentially from DMCD to the semi-hydrogenated methyl 4-
hydroxymethylcyclohexane carboxylate (MHMCC) and subsequent CHDM. Excessive
hydrogenation of CHDM forms the 4-Methyl-1-cyclohexanemethanol (MCHM), and even
1
3

small percentages of cyclohexane and its derivatives (product d-f in Scheme S1). Inappropriate
hydrogenation of MHMCC could also result in the direct removal of one side hydroxyl group
and even –CH
selectivity of the CHDM and relevant by-products including the MHMCC and MCHM are
listed in Table 3. The Re /AC sample gives only 41% conversion of DMCD with the semi-
hydrogenated MHMCC as the major by-product. The incremental content of Re (from 1% to
%) leads to the further hydrogenation of both the ester groups in DMCD to give the final diol
CHDM. The DMCD conversion and CHDM selectivity increases steadily with the Re content.
3% conversion of DMCD and 58% CHDM selectivity can be reached using the Re /AC
3
to form cyclohexanecarboxylic acid (a-c). For clearer comparison, the
1
5
-
9
5
catalyst. Excessive Re decoration results in stable DMCD conversion around 95% but further
transformation of CHDM into other by-products.
Table 3 Catalytic properties of various Re catalysts in the selective hydrogenation of DMCD.a
Conv. (%)
Selectivity of typical products (%)b
Entry Catalyst
P (MPa)
DMCD
41
CHDM
15
MHMCC
MCHM
1
2
3
4
5
6
7
8
9
Re
Re
Re
Re
Re
Re
Re
Re
Re
1
3
5
7
9
5
5
5
5
/AC
8
0
41
29
10
12
11
3
/AC
8
75
32
3
/AC
8
93
58
9
/AC
8
91
53
12
6
/AC
8
92
49
/AC-HO
/AC-N
/AC-H
/AC
10
10
10
10
81
23
25
25
53
79
26
8
15
10
1
99
67
1
20
a
o
Reaction conditions: reaction under H
2
pressure of 8 MPa or 10 MPa at 220 C for 10 h, weight ratio of
1
4

DMCD/catalyst = 4:1. DMCD: Dimethyl 1, 4-cyclohexanedicarboxylate; MHMCC: methyl 4-
hydroxymethylcyclohexane carboxylate; CHDM: 1, 4-cyclohexanedimethanol; MCHM: 4-Methyl-1-
cyclohexanemethanol. b Other detected by-products and the possible formation pathway is illustrated in
Scheme S1.
The influences of reaction conditions have been studied by using Re /AC as a typical
5
sample (Figure 3). Both the reaction temperature and pressure affect the distribution of
products. Lower temperature causes the direct formation of semi-hydrogenated MHMCC with
only one side of ester group transferred into alcohol. The final yield of diol (CHDM) increases
o
volcanically with the reaction temperature and reaches its maximum around 220 C. Further
raising of temperature leads to the formation of MCHM due to the excessive hydrogenation
of one -CH
2
OH into -CH
3
. The variation of reaction pressure performs similar trend on the
o
distribution of products. Reaction condition under 10 MPa H
2
at 220 C gives the better
selectivity of diol at 66% with nearly complete conversion of diester. It is visible that MHMCC
and CHDM are formed as the initial product within the first 4 h. CHDM is accumulated with
the prolonged reaction time, which should be transformed from the MHMCC.
1
5

Figure 3. Catalytic performances of Re /AC for the selective hydrogenation of dimethyl 1, 4-
5
cyclohexanedicarboxylate (DMCD) as a function of reaction temperature (A), reaction pressure (B),
reaction time (C), and catalyst dosage (D). MHMCC: methyl 4-hydroxymethylcyclohexane carboxylate;
CHDM: 1, 4-cyclohexanedimethanol; MCHM: 4-Methyl-1-cyclohexanemethanol. Conditions: reaction
o
under H
2
pressure of 8 MPa at 220 C for 10 h, weight ratio of DMCD/catalyst = 4:1, if not specified.
For better understanding the hydrogenation process, the conversion of DMCD and
selectivity of products as a function of the catalyst dosage are shown in Figure 3D. The
decrease of catalyst dosage results in the higher amount of semi-hydrogenated MHMCC, due
to the insufficient hydrogenation ability by the less amount of catalyst. Whilst excessive
dosage of catalyst (weight ratio of reactant/catalyst < 4:1) causes the drastic hydrogenation
process to form excessively hydrogenated MCHM with -CH . It is interesting that although
3
the higher dosage of catalyst restrains the formation of other complex products, the yield of
DMCD could not be further facilitated. In addition, Re species decorated carbon with different
pre-treatments have also been synthesized and tested for comparison (Table 2, entry 6-9),
1
6

including the H
2
thermal treatment (Re
5
/AC-H), H
2
O
2
treatment (Re
5
/AC-HO), HNO
3
treatment (Re /AC-N), and H
5
2
/HNO bi-functionalization (Re
3
5
/AC). It is remarkable that the
pre-treatments of activated carbon result in the distinctive catalytic behaviors. The Re /AC-H
5
o
displays only 15% conversion of diester (DMCD) at 220 C for 10 h. However, semi-
hydrogenated MHMCC is produced as the main by-product and the selectivity of diol is as
low as 10%. An additional by-product (product g in Scheme S1) was also detected
occasionally in the reaction solution, mainly caused by the condensation between DMCD and
isopropanol, which is only observed by using Re
5
/AC-H. The Re
5
/AC-HO and Re
5
/AC-N
samples have nearly identical catalytic performances with 81% and 79% conversion of DMCD.
The selectivity of CHDM are 23% and 26% by Re
5
/AC-HO and Re
5
/AC-N, respectively. It
behaved as available
was reported that H aqueous solution and concentrated HNO
2
O
2
3
oxidants to oxidize carbon materials and abundant oxygenated groups were easily generated
30, 31]. Based on the bi-functionalization by both H thermal treatment and subsequent HNO
[
2
3
5
oxidation, the Re /AC sample largely facilitates the formation of the target diol product (yield
o
of 66%) with complete conversion of DMCD at 220 C in 10 MPa H
2
.
3
.3 Phase analysis and surface properties of rhenium catalysts
It is until now suspected that two major elements essential for the catalytic performance
and selectivity of diol by the series of Re catalysts: i) The rhenium sites with moderate ability
to activate, adsorb, or dissociate both the H and reactant molecules; and ii) The possible roles
2
of surface functionalization and oxygenated groups to modulate the surface distribution of
rhenium in Re/AC samples. It is straightforward that the surface chemistry of each sample
1
7

should be relevant with their catalytic activities. Herein, XRD is performed subsequently to
determine the possible particle sizes and crystal phases in different catalysts (Figure 4), in
order to rule out other complex surface and structure parameters. Two broad diffraction peaks
o
o
at 20-30 and 40-50 related to the (002) and (100) crystal faces of amorphous carbon can be
observed in all the samples [32, 33]. After the decoration of Ru species, there is no observation
of Re-containing species, independent of the Re content or the pretreatment. Considering the
relatively high amount of Re species, especially in Re
7
/AC and Re /AC, no great accumulation
9
or aggregation is formed by Re species. However, the peak intensity of amorphous carbon
largely declines with the raising metal loadings, indicating the degree of aromatization and the
aromatic planar molecular size were influenced by metal decoration. Re species may change
the surface of activated carbon by surface decoration or even inserting, which may split the
support into small and randomly oriented sheets [34, 35].
Figure 4. XRD patterns of Re /AC-y samples with different Re loadings and pretreatments of activated
x
carbon.
The easy formation of small ReO
x
particles not only change the micro-structure of carbon,
but may also alter the surface chemical environment of the catalysts. In order to analyze the
1
8

surface acidity that may be generated by the carbon surface and the Re decoration, NH -TPD
3
are further carried out. One intensive and broad desorption peak appears with its maximum
o
around 600 C (not shown), mainly due to the reaction between C and NH
3
to form one sharp
o
desorption peak. Overlapped desorption peaks can be observed in the range of 100-400 C, as
enlarged in Figure 5A-B. Pure carbon support displays two desorption peaks maximized at
o
o
1
10 C and 270 C, corresponding to the weak acid sites and medium strong acid sites [36, 37].
The surface acidity of the activated carbon might be from the surface functionalized
oxygenated groups such as –COOH and C-O(H) groups [38, 39]. After Re decoration, the
distribution of acid sites largely varies. In general, metal decoration in Re /AC leads to the
1
weakening of both the weak and medium strong acid sites. Formation of new acid sites at
o
about 350 C are observed. As the Re content increases, the peak location of the medium strong
o
o
acid sites shifts from 350 C to 280 C with much intensified peak area. Whilst the desorption
temperature of the weak acid sites is not obviously changed, confirming the unchanged type
of weak acid sites. The series of Re /AC-y samples (Figure 5B) with different pretreatments
5
display nearly identical location and peak area of the desorption peaks, unless the much
o
delayed desorption peak at 300-350 C in Re
5
/AC-H. It is also necessary to mention that only
o
Re decorated carbon materials have the methanation peak at about 600 C (Figure S1 in
supporting information), whilst activated carbon does not display any observable peak in the
o
range of 400-700 C. Rhenium decoration facilitated the activation of carbon surface. It is also
visualized that the surface acidity of rhenium catalysts, or more accurately the medium strong
acid sites, is directly correlated with the surface dispersed rhenium species.
1
9

Figure 5. NH
3
-TPD profiles (A-B) and Raman spectra (C) of different Re /AC-y samples.
x
In fact, the chemical etching may also happen during pretreatment and Re decoration,
resulting in the modified surface micro-structure and can be revealed by the Raman spectra
-1
(
Figure 5C) [40]. In the wavenumber range of 1800-1000 cm , two broad peaks are displayed
-1
at 1570 and 1340 cm , representing the structural defects and partially disordered structures
2
2
of the sp domains (D band) and the E2g vibration mode of sp C domains corresponding to
the graphitization (G band), respectively. The intensity value of the two peaks (I
D
/I ) suggests
G
3
the percentage of sp hybridized carbon atoms in the framework of carbon, viewed as a sign
of the defect degree. Compared to the I
D
/I value of the pure activated carbon (1.45), the
G
related value after rhenium decoration is in the range of 1.8-2.1, and the calculated value of
defect density is listed in Table 1. It is generally consistent with reports that metal doping or
inserting may largely lead to the formation of structure defects [40, 41], which also supports
the assumption of XRD results. However, as the Re loading increases, the I
D
/I
G
decreases from
2
.09 to 1.89. It is inferred that rhenium decoration results in changes on the carbon surface,
and imperfect surface defects are generated surrounding small ReO particles at lower Re
x
content. Whilst excessive ReO particles are formed at higher Re content, which may cause
x
2
0

the sinking or blocking of surface instead of regional defects. On the contrary, the defect
structure is not greatly influenced by different pretreatment of activated carbon only if the
decorated Re content was similar (5 wt%) in this work.
Oxygenated groups may simultaneously generated with surface defects during the
synthesis and pretreatment [42, 43]. And the XPS spectra in the O1s region is shown in Figure
6
. The broad O1s peak can be deconvoluted into three overlapped peaks. The peaks at binding
energy of 533.5 eV, 532.5 eV and 531.0 eV are corresponding to C-O(H), O-C=O, and C=O
of carbon, respectively [41, 44]. The atomic ratio between O/C of different catalysts varies in
the range of 7.0%-9.5% (Table 2), slightly raises with the incremental Re loadings. The lattice
oxygen in the as-formed ReO species also contributes to the distribution of surface oxygen
x
species at 531.5-530.0 eV, overlapping with the C=O species of carbon. It is also remarkable
that both the percentages of C-O and O-C=O species obviously decrease along with the Re
content. Considering the weakening of acid sites on carbon and the increased value of I
after Re decoration, it is rational that ReO particles are generated occupying the surface
oxygenated groups, especially the C-O and O-C=O groups. The new generated lattice oxygen
in ReO replenishes the raising percentage of oxygen species at 531.0 eV from 21% to 56%
as the Re loading raises from 3% to 9%. The different types of pretreatments also cause the
variation of the O/C ratio. Pretreatments of activated carbon by H and HNO lead to similar
O/C ratio and distribution of surface oxygen species. Using activated carbon after H
treatment, the Re
groups, which partially accounts for the much intensified desorption peaks in NH
Re /AC-H, compare with other samples.
/I
D G
x
x
2
O
2
3
2
thermal
5
/AC-H sample have 51% O-C=O coming from the anhydride and carboxyl
3
-TPD of
5
2
1

Figure 6. XPS spectra in the O1s region of Re
N (E), and Re /AC-H (F) samples.
3
/AC (A), Re
5
/AC (B), Re
9
/AC (C), Re
5
/AC-HO (D), Re /AC-
5
5
For better confirmation of the metal surface distribution and other surface information,
the STEM and high-resolution TEM of Re /AC are illustrated in Figure 7A-B. The shining
spots in Figure 7A suggest the surface distribution of ReO particles. Small Re particles are
5
x
formed on the surface of activated carbon. Two different location in the TEM images are
amplified as shown in Figure 7B-I and B-II. The high-resolution TEM images illustrate the
distribution of ellipsoidal particles on the surface of activated carbon. The interplanar distances
of 2.24 Å, 2.09 Å and 1.90 Å are measured and are corresponding to the (002), (201) and (111)
2
2

o
o
crystal faces of ReO
Figure 7B-I further confirms the existence of ReO
particles about 2 nm are also observed on the surface of activated carbon, as pointed by the
dotted cycle in Figure 7B-II. It is inferred that small ReO particles are readily formed in the
Re /AC sample and are possibly melted to form the ellipsoidal particles, due to the surface
orientation effect during thermal treatment. The ReO (111) crystal faces are more frequently
3
. The intersection angle of 65 and 53 between different crystal faces in
3
phases. Very small and spherical ReO
3
3
5
3
exposed with extensive surface platform in the special ellipsoidal particles. Interestingly,
terrace sites are likely to be generated and exposed due to the interlaced (111) and (002) crystal
faces as shown by the arrows in Figure 7B-I.
Figure 7. STEM image (A), high-resolution TEM images (B), and CO chemisorption by pulse (C) of typical
Re
5
/AC sample, and the relationship between specific rate and Re content (D), B-I and B-II are the enlarged
composition. And the
parts in the dotted box of B. The arrows in B-I point out the terrace sites of ReO
3
dotted cycles in B-II illustrate the well-formed small particles around 2 nm. The red balls in D show the
o
specific rates of Re
Re
x
/AC at 220 C in 8 MPa H
2
, and green balls suggest the specific rates of different
o
5
/AC-y samples at 220 C in 10 MPa H
2
.
No matter how, the stability of the monometallic Re/AC still has to be enhanced. The
2
3

catalytic performance of the Re /AC sample is not satisfied as the DMCD conversion
5
decreases continuously during the four-time reaction cycles (Figure S2). The selectivity of
CHDM also decreases during the third time of reaction. It is suggested that the re-used sample
becomes incompetent to activate the diester and to hydrogenate the intermediate in depth,
mainly due to the variation of its surface composition or the structure. The TEM image of used
Re /AC sample is displayed in Figure S3. It is observable that particles on the surface of AC
5
seem to be surrounded by the rough surface of amorphous carbon, which is possibly formed
during the reaction and causes the coking on the catalyst. The failed exposure of the active
ReO
x
species is responsible for the much watered-down catalytic behavior. Promoted Re/AC
sites should be more reliable for the reaction.
catalyst with availably and stably exposed ReO
x
3
.4 On the acting roles of rhenium decoration
Until now, it is generally clear that Re decoration on activated carbon is responsible for
not only the active metal sites, but also for the great modifications on both the surface chemical
environment and micro-structure. Combining the results from XPS and NH -TPD, the
decoration of Re is mostly accomplished by the rhenium species occupying the surface
oxygenated groups on carbon during surface functionalization. The new generated ReO
3
x
species changes the porous structure and the surface defect density of carbon. The specific
rates according to the exposed Re active sites are evaluated based on the CO chemisorption as
listed in Table 1 and Figure 7C-D. Unlike the exponentially increased CHDM yield, the
specific rate of Re
x
/AC samples losses as a function of the Re content. The Re /AC sample
1
2
−ꢁ
푅푒
−ꢁ
displays the much exaggerated specific rate of 11.9 × 10 푚푚표푙퐷푀퐶퐷 ∙ ꢀ ∙ ℎ , indicating
2
4

its superior utilization of each exposed metal atom compared to the other samples with higher
Re contents. The specific rates of Re
5
/AC are calculated to be 7.7 × 102 and 9.5 ×
02 푚푚표푙퐷푀퐶퐷 ∙ ꢀ ∙ ℎ , under 8 MPa and 10 MPa H
−ꢁ
−ꢁ
at 220 C, respectively. It has to be
o
1
2
푅푒
noted that only Cr based catalysts have been reported on an industrial scale. Thus the current
result based on Re catalyst has been compared with the reported literature (Table S1).
Although Cu based catalysts displayed higher CHDM yield compared to the current and
reported work based on precious metals, the Cu loadings were extremely high [45, 46]. The
calculated specific rate of the reported Cu/ZnO/ZrO
2
catalyst was only 16.8 mmol/gCu/h at 220
oC in 8 MPa H
[47]. Other bimetallic Ru-Sn samples and even trimetallic samples were
2
reported for the selective hydrogenation process to give CHDM [48, 49]. However, higher
reaction temperature and excessive use of precious metals were necessary for better activity.
No matter how, Re based catalyst can be potentially applied to the selective production of
CHDM with further constructive modification.
As a direct result of the lower Re content, finer ReO particles with Re species in reduced
x
0
4+
states (Re and Re ) are also facilitated to be dispersed on the surface only if the Re loading
was lowered to 5 wt% and even less. Such phenomenon suggests that the selective
hydrogenation of DMCD might be structure-sensitive. No matter how, although the ReO
x
particles slightly accumulate due to the shortened distances between neighboring particles,
there is no sign of great sintering at high Re loading as revealed by XRD and TEM. Herein, it
is more rational to propose the surface dispersion of exposed active Re sites (Table 1), instead
of the ReO
x
particle sizes, that are adjusted controllably by the Re decoration content.
Appropriately enhanced Re content is preferential to produce selectively the target CHDM
2
5

with enough power. The higher surface dispersion of ReO inducing the better redox property
x
of the catalyst and the more observable interaction between Re and carbon. Electron
transformation has been confirmed from Re species to the neighboring C atoms to give the
availably exposed electron-deficient Re sites. In the typical Re /AC sample, ellipsoidal
5
particles exist with terrace sites at the cross interfaces of (111) and (002) faces. Large amount
of medium strong acid sites are constructed simultaneously, which positively accelerate
relevant with the catalytic performance.
4
. Conclusion
A series of Re catalysts on activated carbon have been synthesized and used for the
selective hydrogenation of diester to diol. The initial influences on both the Re surface
dispersion and pretreatment of activated carbon have been investigated. ReO particles were
readily generated by occupying the surface oxygenated groups on carbon (mostly C-O(H) and
H)O-C=O groups). Fine ReO particles were dispersed on the carbon surface independent of
x
(
x
the synthesis conditions, mainly due to the initial electron-positive property and its electron
transferred from Re to C. It ascribed subsequently for the much facilitated reducibility of the
catalyst and the maintenance of Re6+ even after H
reduction. ReO species as electron
2
x
acceptor could further act as appropriate acid sites and adsorb the reactant more readily. Re
species in the reduced states could still be maintained under appropriate Re surface dispersion.
The moderate hydrogenation ability of Re based catalyst restrained the excessive
hydrogenation of CHDM to other by-products. The complete conversion of diester and 66%
yield of diol can be reached with the specific rate of 9.5 × 10 푚푚표푙퐷푀퐶퐷 ∙ ꢀ ∙ ℎ−ꢁ at 220
2
−ꢁ
푅푒
2
6

oC under 10 MPa H
.
2
Author Statement
Jingjie Luo: Conceptualization, Investigation, Writing - Original Draft. Enhui Qu: Data
curation, Investigation. Yixue Zhou: Validation, Yanan Dong: Formal analysis.
Changhai Liang: Writing- Reviewing and Editing, Supervision, Project administration,
Funding acquisition,
Declaration of interests
The authors declare that they have no known competing financial interests or personal
relationships that could have appeared to influence the work reported in this paper.
Acknowledgements
This work was supported by the National Natural Science Foundation of China (21573031,
2
2
1978031 & 21802015), and Program for Excellent Talents in Dalian City (2016RD09,
016RD04). The authors acknowledge Dr. Zhanglong Guo for the experimental assistance of
STEM and HRTEM.
References
[
1] J. Wang, S. Mahmud, X. Zhang, J. Zhu, Z. Shen, X. Liu, Biobased amorphous polyesters
with high T : Trade-off between rigid and flexible cyclic diols. Acs Sustain. Chem. Eng. 7
2019) 6401-6411. https://doi.org/10.1021/acssuschemeng.9b00285.
g
(
[
2] S. Hahm, J.-S. Kim, H. Yun, J.H. Park, R.A. Letteri, B.J. Kim, Bench-scale synthesis and
2
7

characterization of biodegradable aliphatic–aromatic random copolymers with 1,4-
cyclohexanedimethanol units toward sustainable packaging applications. Acs Sustain. Chem.
Eng. 7 (2019) 4734-4743. https://doi.org/10.1021/acssuschemeng.8b04720.
[
3] M. Matos, A.F. Sousa, A.J.D. Silvestre, Improving the thermal properties of poly(2,5-
furandicarboxylate)s using cyclohexylene moieties: A comparative study. Macromol. Chem.
Phys. 218 (2017) 1600492. https://doi.org/10.1002/macp.201600492.
[
4] A. Guarneri, V. Cutifani, M. Cespugli, A. Pellis, R. Vassallo, F. Asaro, C. Ebert, L.
Gardossi, Functionalization of enzymatically synthesized rigid poly(itaconate)s via post-
polymerization Aza-Michael addition of primary amines. Adv. Synth. Catal. 361 (2019) 2559-
2
573. https://doi.org/10.1002/adsc.201900055.
[
5] G. Lewandowski, A. Wroblewska, E. Milchert, Synthesis of 1,4-cyclohexanedimethanol
by hydrogenation of dimethyl terephthalate and its application as a substrate in syntheses of
polyesters. Polimery-W 52 (2007) 39-43. https://doi.org/10.14314/polimery.2007.039.
[
6] S.Y. Zhang, Q. Hu, G.L. Fan, F. Li, The relationship between the structure and catalytic
performance Cu/ZnO/ZrO catalysts for hydrogenation of dimethyl 1,4-cyclohexane
dicarboxylate. Catal. Commun. 39 (2013) 96-101.
http://doi.org/10.1016/j.catcom.2013.05.011.
2
[
7] J.M. Thomas, B.F.G. Johnson, R. Raja, G. Sankar, P.A. Midgley, High-performance
nanocatalysts for single-step hydrogenations. Accounts Chem. Res. 36 (2003) 20-30.
http://doi.org/ 10.1002/ chin.200316294.
[
8] R. Raja, T. Khimyak, J.M. Thomas, S. Hermans, B.F.G. Johnson, Single-step, highly
active, and highly selective nanoparticle catalysts for the hydrogenation of key organic
compounds. Angew. Chem. Int. Edit. 240 (2001) 4638-4642. https://doi.org/10.1002/1521-
3
773(20011217)40:24<4638::AID-ANIE4638>3.0.CO;2-W.
[
9] Q. Hu, G.L. Fan, S.Y. Zhang, L. Yang, F. Li, Gas phase hydrogenation of dimethyl-1,4-
cyclohexane dicarboxylate over highly dispersed and stable supported copper-based catalysts.
J. Mol. Catal. a-Chem. 397 (2015) 134-141. http://doi.org/10.1016/j.molcata.2014.10.032.
2
8

[
10]J. Zhang, C. Li, X. Chen, Y. Chen, L. Zhang, B. Zhang, C. Liang, Promotional effects of
magnesia on catalytic performance of Pt/SiO in hydrogenolysis of dibenzofuran. J. Catal. 371
2019) 346-356. https://doi.org/10.1016/j.jcat.2019.02.017.
2
(
[
11]J.M. Keels, X. Chen, S. Karakalos, C. Liang, J.R. Monnier, J.R. Regalbuto, Aqueous-
phase hydrogenation of succinic acid using bimetallic Ir–Re/C catalysts prepared by strong
electrostatic
adsorption.
Acs
Catal.
8
(2018)
6486-6494.
https://doi.org/10.1021/acscatal.8b01006.
[
12]X.X. Dong, B.S. Jin, Y.Q. Sun, L. Yu, Urban gas production from low H
2
/CO biogas using
Re-promoted Ni catalysts supported on modified manganese sand. Fuel 220 (2018) 60-71.
http://doi.org/10.1016/j.fuel.2018.01.128.
[
13]X. Jin, P.S. Thapa, B. Subramaniam, R.V. Chaudhari, Kinetic modeling of sorbitol
hydrogenolysis over bimetallic RuRe/C Catalyst. Acs Sustain. Chem. Eng. 4 (2016) 6037-
047. http://doi.org/10.1021/acssuschemeng.6b01346.
6
[
14]T. Ayvali, P.F. Fazzini, P. Lecante, A. Mayoral, K. Philippot, B. Chaudret, Control of
reactivity through chemical order in very small RuRe nanoparticles. Dalton T. 46 (2017)
5070-15079. http://doi.org/10.1039/C7DT02287E.
1
[
15]X. Di, C. Li, B.S. Zhang, J. Qi, W.Z. Li, D.S. Su, C.H. Liang, Role of Re and Ru in Re-
Ru/C bimetallic catalysts for the aqueous hydrogenation of succinic acid. Ind. Eng. Chem. Res.
6 (2017) 4672-4683. http://doi.org/10.1021/acs.iecr.6b04875.
5
[
16]X. Di, C. Li, G. Lafaye, C. Especel, F. Epron, C.H. Liang, Influence of Re-M interactions
in Re-M/C bimetallic catalysts prepared by a microwave-assisted thermolytic method on
aqueous-phase hydrogenation of succinic acid. Catal. Sci. Technol. 7 (2017) 5212-5223.
http://dio.org/10.1039/c7cy01039g.
[
17]Z.-Y. Li, M.S. Akhtar, D.-H. Kwak, O.B. Yang, Improvement in the surface properties of
activated carbon via steam pretreatment for high performance supercapacitors, Appl. Surf. Sci.
04 (2017) 88-93. https://doi.org/10.1016/j.apsusc.2017.01.238.
4
2
9